IJUC HomeIssue Contents

Preface
Ella Gale and Ioannis Georgilas

The contributions of this special issue were presented at the firstWorkshop on Unconventional Approaches for Robotics Automation and Control Inspired by Nature (UARACIN) which was supported by the IEEE RAS Technical Committee on Biorobotics and was held as part of the International Conference on Robotics and Automation (ICRA 2013) on 10th May in Karlsruhe, Germany. The one-day workshop included three keynotes given by leaders in the field: Andrew Adamatzky on Physarum Polycephalum and its applications to computing, Georgios Sirakoulis on Cellular Automata (CA) for the advancement of robotics and a joint talk by Paolo Arena and Roland Strauss on fruit fly brains and modelling them in hardware for autonomous robot control. Eleven short discussion papers were presented covering a wide-range of unconventional computing topics such as memristors, reaction-diffusion computers, cellular automata, and biological computation, with applications to robotics problems such as Simultaneous Localisation And Mapping (SLAM), visual perception, time-perception, wireless sensor networks and control of autonomous exploring robots.

Robotics, automation and control is a field that has been around for many years and one which has had a great impact on science, technology and society as a whole. To build on this, robotic control systems need to be made more resilient to cope with real world conditions. Furthermore, classical approaches to control can not cope with the requirement for massive parallelism that next-generation robots will need. Natural biological and chemical systems are resilient and offer a wealth of ideas for how to solve such problems. Good candidates are the so-called unconventional computing methods, where biological and biologically-inspired methods, primarily from the cellular/bacterial level, are exploited for their natural parallelism to tackle computational complexity, while unconventional hardware, i.e. Physarum, biological computers, BZ reactions, memristors, is used as a novel physical layer exploiting the same natural features. Recent success in biomimetics using unconventional approaches could revolutionise robotic hardware and control. The three papers presented in this special issue demonstrate the utility of unconventional computing approaches for robotics.

Contributions to the special issue

Hey Physarum! Can you perform SLAM? tackles the robotic task of Self Localisation and Mapping (SLAM) using a combined approach inspired by the foraging behaviour of slime mould, Physarum polycephalum, and the computational tool of cellular automata.

Reaction-Diffusion based Computational Model for Autonomous Mobile Robot Exploration of Unknown Environments introduces a Reaction-Diffusion-based model for path-planning of robot locomotion in unknown environments, where decision making is based on the underlying dynamics of the RD system.

Model Accuracy Assessment in Reaction-Diffusion Pattern Formation in Wireless Sensor Networks demonstrates that the use of RD patterns in the design of sensor networks can be used to evaluate the impact of the internode broadcast range and allow an overall assessment of the networks performance.

Ella Gale
Ioannis Georgilas
November, 2013

Full Text (IP)